What IBM’s AI Experiment Can Teach Marketing Leaders About Building Better Teams

In 2023, IBM cut thousands of HR roles and handed the work to AI. Its platform, AskHR, handled 94% of routine tasks, processed over 11.5 million interactions, and helped deliver $3.5 billion in productivity gains. On paper, it was a textbook win.

Then the other 6% started to bite.

Sensitive workplace issues went unresolved. Morale dipped. As Fortune reported, IBM's CHRO acknowledged the early rollout made workers miserable before the company course-corrected. CEO Arvind Krishna later told the Wall Street Journal that while AI had replaced several hundred HR roles, IBM's overall headcount actually grew. The savings were reinvested into software engineers, marketers and salespeople. Roles that require critical thinking and human connection.

This is not a story about IBM getting it wrong. It is a story about what happens when the 6% gets overlooked, and what marketing and sales leaders can learn from it.

The 94% is real. So is the 6%.

If you lead a marketing or sales team, AI is already reshaping your day. Campaign reporting that used to take hours can be done in minutes. Content drafts, lead scoring, customer segmentation, ad optimisation, workflow automation: these tasks are faster and, in many cases, better when AI handles the heavy lifting.

At BFJ, we use AI across our operations. It helps us move faster on data analysis, content production, reporting and campaign execution. We have built prompt libraries, tested agentic workflows and integrated AI into the way our teams work every day.

But here is what I have learned leading teams through this shift: the efficiency gains only matter if you know what to do with them.

IBM’s experience shows that automating the repetitive 94% is the straightforward part.

The harder question, and the one that separates good teams from great ones, is how you handle the remaining 6%.

In marketing and sales, that 6% looks like a strategist recognising that a technically “successful” campaign isn’t actually moving the business forward. It looks like an account manager sensing a client is frustrated before they say it out loud. It looks like a creative director pushing back on safe work because the brief deserves better.

These are not tasks you can automate. They are the product of experience, emotional intelligence and the kind of trust that only builds between real people over time.

The leadership trap

There is a version of AI adoption that is purely about cost reduction. Cut headcount, automate output, report the savings. On a spreadsheet, it works.

In practice, it creates the same gaps IBM discovered: the work that requires judgement, creativity and human connection starts to degrade quietly.

You don’t always notice it until the damage is done, a client leaves, a team member burns out, or the work starts to feel interchangeable with everyone else’s.

The better approach, and the one I encourage every marketing leader to consider, is to treat AI as a way to make your people more effective rather than to replace them. When you free your team from repetitive work, you create space for the thinking that actually drives results: better strategy, deeper client relationships, sharper creative.

In professional services, the quality of your people is your competitive advantage. AI can accelerate what good teams do. It cannot build the trust, the culture or the judgement that makes those teams good in the first place.

What this means in practice

For the marketing and sales leaders reading this, here is where I would start.

Be honest about what AI does well in your business. Audit your team’s time. Identify the tasks that are repetitive, data-heavy and process-driven. That is where AI delivers the clearest return: reporting, segmentation, content production at scale, workflow automation, data hygiene. Invest there first.

Protect the work that requires humans. Strategy, relationship management, creative direction, complex problem-solving: these are the areas where your team’s experience and judgement create the most value. If AI is eating into time spent on these activities rather than freeing time for them, something has gone wrong.

Be transparent with your team. The fastest way to erode trust during an AI rollout is to let people guess what it means for their roles. The teams I have seen navigate this well are the ones where leadership is upfront: here is what we are automating, here is why, and here is how it changes your role for the better.

Reinvest the efficiency gains. This is the lesson from IBM. The savings from automation are only valuable if they go somewhere meaningful: deeper training, better tools, additional headcount in strategic roles, or simply giving your team more time to think. The reinvestment is what turns efficiency into performance.

The real competitive advantage

Every digital marketing agency, every marketing team and every SaaS platform is adopting AI.

The tools will reach parity.

What will not reach parity is the quality of your people, the clarity of your strategy and the trust you have with your clients.

AI is the most powerful tool marketing and sales teams have gained in a generation. But it is still a tool. The teams that win over the next few years will not be the ones that automated the most. They will be the ones that understood what to automate, what to protect and how to lead people through the change with honesty and purpose.

That 6% IBM could not automate? In marketing, it is often the 6% that matters most.